The generation of cellular diversity is essential to the development of multicellular organisms. Understanding the genetic mechanisms that specify cell fates provides insight into diseases, such as cancer, that co-opt normal developmental programs. The Drosophila melanogaster retina has long been used as a paradigm for studying fate specification; the retina consists of ∼750 ommatidia, each containing eight photoreceptors (R1-R8), which adopt unique morphologies and functions. R1-R6s are responsible for motion detection while R7s and R8s are used for color discrimination, analogous to mammalian rods and cones respectively. Within each ommatidium, photoreceptors are recruited in a stereotypical order, the last three, R1, R6, and R7, forming an equivalence group. Notch signaling breaks the equivalence of the R1/R6/R7 precursors: while all three precursors express both Notch receptor and the Notch-ligand Delta, signaling occurs in only one direction, with R1 and R6 (R1/R6) redundantly using Delta to activate Notch in R7. Why is Notch activated only in R7? In Chapter III describe work showing that the timing of Delta expression and Delta-mediated cis-inhibition of Notch generate the unidirectionality of signaling. This work represents the first evidence that cis-inhibition biases the directionality of Notch signaling. How does Notch activation specify R7 fate? In Chapter III I show that Notch acts in a novel branched pathway: it represses the transcription factor seven-up and thereby the R1/R6 fate, and in parallel biases a normally hidden, stochastic fate choice toward the R7 fate. This work suggests that stochastic fate choices can underlie even deterministic specification of fate by extracellular signals. After R7s are specified, they face the binary choice between the pale or yellow subtype; each subtype expresses a different light-sensitive rhodopsin necessary for color vision. This choice is regulated by the stochastic expression of the transcription factor Spineless, but the genes it regulates are unknown. In Chapter IV I show that Spineless represses the transcription factor Runt and thereby the pale subtype and I examine the action of both genes which ensure that R7s are stably specified as either pale or yellow. This dissertation includes published and co-authored material.